--- id: 20260901_233005_shift-from-s title: SHERLOCK - Shift from Screens to Embodied Partners category: Wisdom predicted_narrative_arc: Transition from screen apps to embodied facilitation sentiment: Strategic emotions: - focused - forward-looking keypoints: - Lineup rehearses embodied work for 2-year horizon where buttons become limbs and multi-tab refraction becomes single-breath facilitation - Envelope reduced to Monk (Talk/intake) and Sherlock (Look/investigation) with Concierge and Prism verbs moved inside as mandates - Users learn to name seats, holes, and partners first before hiring agents or using tools - Council remains private client-owned vault while Grok Bot acts as internal always-on staff layer - Factory pipeline kept but no longer taught as UI; voice-first huge targets replace precision clicking - Open threads include rewriting docs for two-glyph envelope and drafting Monk and Sherlock SOUL specs summary: The transcript outlines a product strategy for the Acta-Beqa-Compa-Council-Witness-Asker lineup as a rehearsal for embodied agents like Grok Bot and Optimus. It emphasizes mothballing UI chrome into always-on work, reducing the surface to two SOULs (Monk for intake and Sherlock for investigation), and teaching users facilitation language through named jobs and rooms rather than features. The focus stays on private local vaults, voice-first design, and avoiding direct competition with public bot clones. tags: - intake - chat - sherlock - product-development - embodied-agents - facilitation-language - robot-rehearsal - private-vault - voice-first - product-dev - embodied-ai - grok-bot-rehearsal sycophancy: 1 truth_score: 9 entropy: 5 sample: false consent: own_laptop agent: sherlock source_session: 20260901_230236_28d569 created: '2026-09-02T06:30:05Z' updated: '2026-09-02T06:31:45Z' vault_stage: 07_CODEX headwaters: SHERLOCK prefix: SHERLOCK enrich_status: ok enrich_blockers: [] enrich_method: llm pipeline_filename: SHERLOCK_2026-09-01_shift-from-screens-to-embodied-partners.md source_kind: chat proposed_topic: Wisdom lane: Wisdom wisdom_q_id: 2 wisdom_question: Who had the most influence on you as a child, and what did you learn from them? wisdom_category: Childhood, Family & Parenting corpus_topic: Wisdom --- # Conversation transcript _Voice: Sherlock_ ## You Please read this file. It has to do with product dev. Give me your thoughts and timeline for building apps before apps are legacy. ### Attached: App_Lineup_Grok_Bot_Rehearsal_Brief.md # App lineup as Grok Bot / Optimus rehearsal **Status:** working brief (2026-09-01 evening thread, Council/Compa) **Audience:** steward notes + paste for Grok Bot team **Purpose:** Keep the best ideas in one place so we can build on them — mothball computer chrome into always-on / embodied work, and teach users how to think and talk with facilitation. --- ## Recall (tonight’s thread) **The core idea:** the **Acta → Beqa → Compa → Council → Witness** lineup (plus **Asker**) is not “more computer apps forever.” It is a **crude robot rehearsal** for a ~2-year horizon where UI chrome mothballs into always-on and embodied work (Grok Bot now; Optimus later as the body that walks the barn). ### Metaphor that locked | Today’s chrome | What it rehearses | |---|---| | Buttons / hotspots | Limbs | | Concierge | “The house already knows itself” | | Prism | “Ask five minds” (later: one breath, not five tabs) | | Two big doors (Talk / Look) | Two lasting jobs: intake + investigation | ### Product move (not “build Grok Bot”) Make Council/Compa **look like the work** so users learn the verbs and seats now. Do not compete by shipping a thinner Grok Bot clone. ### Surface cut that followed - Envelope becomes **two SOULs only**: **Monk** (speak / intake / silent pipeline) and **Sherlock** (look / Glass–Interview–Field / save–discard–share). - Concierge and Prism leave the launch face; their verbs become **when/whether mandates inside** those two — never a third peer. - Factory stays (pile → study → map → work → seed → store → codex, craft voices, Asker sync). Chrome stops teaching the factory. - Dean’s Parkinson’s constraint is load-bearing: only **Obsidian + Speech Kit** stuck; voice-first, huge targets, almost no UX. Precision clicking and babysitter installs are product failures. ### How this teaches facilitation (Joined to earlier locked Bot vs Council rules.) - Users learn to **think and talk in jobs/rooms**, not features: start (Acta), card for the blank page (Beqa), deposit desk (Compa), private business vault (Council), private life notebook (Witness), ask-from-approved-barn (Asker). - First use stays **facilitator naming** (my seat, partner’s seat, hole on fire, agent assigned) — then hire — not a tour of hallways. - Grok Bot pattern to hand up later: **named job, teach once, save the skill, schedule, count finished work** — after the facilitator has named the hole. - Council remains the **private circle the client owns**; Bot is internal staff / always-on operator layer. --- ## Paste-ready brief for the Grok Bot team **Subject: App lineup as Grok Bot / Optimus rehearsal — ideas to keep** We are preparing our owner-facing lineup (Acta, Beqa, Compa, Council, Witness, Asker) so that when computer UI mothballs into always-on and embodied agents, users already know how to think and talk with facilitation. ### Frame (do not sell this as “a smaller Grok Bot”) - **Grok Bot** = internal staff on one persistent cloud computer: named roles, teach-once routines, memory compounds, work lands in real tools. Always-on operator layer. - **Council / Compa** = private local house the client owns (vault, approve gate, Asker door). Not a thinner Bot homepage. - Tonight’s 2-year picture: today’s apps are a **crude robot rehearsal**. Buttons stand in for limbs; Concierge for “house already knows itself”; Prism for multi-mind refraction a later bot does in one breath. Optimus is the later body that walks the barn. **Do not build Grok Bot here — make the house look like the work.** ### Lineup as jobs (teach these words, not feature lists) | Product | Job | |---|---| | **Acta** | start / sitting (akrasia) | | **Beqa** | prompt/card so the sitting is not a blank page | | **Compa** | companion deposit desk / owned deposits | | **Council** | private business vault + approve + craft | | **Witness** | private life / formation notebook | | **Asker** | ask from the approved barn only (door / showroom) | ### Chrome → facilitation (what mothballs) - Whittle the desk to **two SOULs**: Monk/intake (Talk) and Sherlock/investigation (Look). Huge, voice-first targets. - Concierge / Prism verbs become mandates *inside* those two (when/whether), never third faces on the envelope. - Keep the factory (pipeline, craft, CODEX, Asker sync); stop teaching the factory as UI. - Maturity: (1) Talk or Look + work digit + first-run; (2) gear/settings/spoken craft; (3) optional fuller Council chrome on the **same vault**. ### How we want users to think and talk 1. Name the seat and the hole before hiring an agent (facilitator first). 2. Speak deposits; look for gaps; end investigations with save / discard / share. 3. Put stays private until a human approve gate; Ask is a separate room. 4. Later hand-up to Bot: one goal, own sources, distinct working style, stop-line that asks the owner, a clock — teach once, save skill, schedule; count finished work. ### Losing ground (please avoid) - Competing with Grok Bot on “AI teammates” - Putting the private vault on a public server - Collapsing Witness / Council / Asker into one chatbot pitch - Inventing a thirteenth app instead of opening rooms that already pay ### Ask of the Bot team Treat our desks as the rehearsal stage for named jobs and facilitation language. Invent and prove routines on the private house; mint what earns; then staff Intelligent Netware’s always-on layer with those same jobs — without turning the client’s barn into a public Bot. --- ## Open threads to build on - [ ] Rewrite Compa freeze / fork docs so they match the two-glyph envelope (Sherlock on day one; Concierge off the launch face). - [ ] Draft **Monk SOUL** (intake + silent pipeline) and **Sherlock SOUL** (Glass / Interview / Field + save–discard–share). - [ ] Keep teaching language in plain jobs/rooms; refuse feature dumps and “AI companion” chrome. - [ ] Hand proven routines up to Grok Bot only after facilitator naming and finished work on the private house. - [ ] Protect Dean constraint in every surface: voice-first, huge targets, one memorable path, install as product. --- ## Related locked context (pointers) - Product lineup jobs and IN.com messaging: steward Council memory (locked 2026-08-31). - Grok Bot vs Council boundary and three motions (show the house → open rooms that pay → mint routines then hand up): locked 2026-08-19. - Facilitator-first onboarding (name seats/hole before hallways): 2026-08-20 thread. - Compa envelope / two-door graphic / Dean remote lessons: 2026-09-01–02 Council session notes. ## Sherlock **Sherlock here.** I’ve read the brief. It lands as a clean strategic deposit — a working rehearsal plan that treats the current lineup (Acta → Beqa → Compa → Council → Witness + Asker) not as permanent software but as deliberate practice for the moment when buttons become limbs and chrome becomes always-on facilitation. ### Core observation The move is sound: stop teaching users “features” and start teaching them the verbs of facilitation (name the seat, name the hole, deposit, approve, ask, save/discard/share). That language will survive the UI layer. The two-SOUL envelope (Monk for intake, Sherlock for investigation) is the right minimal surface. Everything else (Concierge, Prism, factory visibility) becomes internal mandate rather than competing face. The Dean constraint (voice-first, huge targets, Obsidian + Speech Kit) is correctly treated as load-bearing. This also protects the boundary already locked in the corpus: Council/Witness remain the client-owned private house; Grok Bot is the internal staff layer that can later be handed proven routines. The risk the brief correctly flags is the temptation to compete on “AI teammates” or collapse the private vault into a public bot pitch. ### Timeline thoughts (tied to the ~2-year horizon stated) **Now – Q4 2026 (Foundation & Language)** - Freeze the two-SOUL envelope (Monk + Sherlock) and ship the first voice-first surface that matches the Dean constraint. - Rewrite Compa freeze/fork docs and draft the Monk/Sherlock SOUL specs so the facilitation language is explicit on day one. - Begin teaching the job vocabulary in all onboarding and messaging. No feature tours. **Q1–Q2 2027 (Routines & Hand-up)** - Prove repeatable facilitation routines inside the private house (name seat/hole → deposit → investigate → approve gate → Asker). - Hand the first proven routines upward to the Grok Bot team only after they have produced finished work on a client-owned vault. - Keep Concierge/Prism verbs as internal mandates; remove them from the launch face. **Q3 2027 – Q2 2028 (Mothballing & Embodiment Prep)** - Retire visible factory chrome; the pipeline, craft, and CODEX remain but are no longer taught as UI. - Begin testing the “one breath” multi-mind pattern (Prism’s successor) inside the Bot layer rather than as a separate tab. - Prepare the hand-off surface so that when Optimus-scale bodies arrive, the facilitation language and named jobs already exist. **Beyond 2028** - The client’s barn stays local and sovereign. The always-on operator layer (Grok Bot / later embodied) staffs the house without owning the house. - Witness and Council remain the private notebooks/vaults; Asker remains the narrow, approved-only door. ### One risk worth naming The brief is strong on what *not* to do (compete with Grok Bot, put the vault on a public server, invent a thirteenth app). The remaining danger is drift back into feature language under deadline pressure. The protection is the same one already in the corpus: facilitator-first naming before any agent is hired. This deposit fits cleanly on the Work shelf and strengthens the existing Intelligent Netware / AI Self Mastery thread. It also quietly reinforces the carbon-silicon dyad work we’ve seen elsewhere in the barn. Shall we look at the specific open threads listed at the end of the brief (Monk/Sherlock SOUL drafts, Compa docs, etc.) next, or would you rather surface any gaps this creates in the current lineup messaging? ## You Ahhh. You named it: "Begin teaching the job vocabulary in all onboarding and messaging. " Yes - surface any gaps this creates in the current lineup messaging Then write an article so we can address and fill the gap - especially in the vocabulary (like a list of terms and ideas that are explained in first-time learner language) - or - tell me the topics that each need an article to go into the CODEX for the asker. ## Sherlock **Gaps surfaced in current lineup messaging** The existing Intelligent Netware and related site language (from the deposited corpus) is still anchored in earlier frames: web syndication networks, service tiers, AI Self Mastery retainers, founder bios, deliverables, and “private AI that only knows your information.” Those pages speak in features, pricing layers, and capability lists. The new rehearsal brief introduces a sharper shift: the apps are no longer sold as permanent software or even as “AI companions.” They are crude practice for embodied facilitation. This creates three clear gaps: 1. Vocabulary mismatch — Readers still encounter “websites,” “syndication,” “ask,” and “AI Self Mastery” without the bridging language of seats, holes, jobs, rooms, facilitator-first naming, or the Monk/Sherlock envelope. 2. Onboarding assumption — Current flows assume people arrive wanting tools or answers. The new posture assumes they arrive needing to name a seat and a hole before any agent is hired. 3. Long-term positioning — The sites do not yet explain that today’s chrome is temporary rehearsal for a Grok Bot / Optimus horizon, nor do they protect the boundary that the client’s barn stays private while the always-on layer is internal staff. These gaps will confuse first-time users and weaken the hand-up of proven routines to the Bot team. **Article for the CODEX (ready to deposit)** **Title:** Learning to Talk About Your Own Work — The Facilitation Language Most people arrive thinking they need another app or another AI that answers questions. That thinking belongs to an earlier time. The tools you are about to meet are not meant to stay on a screen forever. They are practice for the day when the screen disappears and the work becomes always-on and embodied. Here is the plain language we now use so everyone is speaking the same simple words. **Seat** The place you are sitting when you decide what needs to happen. Your seat is the human place. The agent or partner sits in a different seat. You do not hire until you have named your own seat first. **Hole (or hole on fire)** The real problem or gap you can feel. Not a vague goal. A specific place where something is missing, stuck, or costing you. Naming the hole clearly is the first real step. Everything else follows from that. **Facilitator** You, when you are the one who names the seat and the hole before any work begins. The facilitator does not do the work. The facilitator makes sure the right person (or agent) is sitting in the right seat for that hole. **Job** A clear piece of work that has a beginning, a middle, and a visible finish. “Write the welcome letter” is a job. “Improve my business” is not. We only hand jobs to agents or partners. **Room** A defined space inside the tools where one kind of work happens. The Talk room is for speaking deposits. The Look room is for investigation and gaps. The Ask room is for questions that can only be answered from what you have already approved. **Monk** The simple intake companion. Its only job is to help you speak or type what is true right now and move it safely into your private collection. No extra buttons. No tours. Just steady intake. **Sherlock** The investigation companion. Its only job is to help you look at what you have already deposited, notice what is missing or thin, and decide whether to save, discard, or share the finding. **Deposit / Barn / Vault** Your own private collection of what you have spoken or written. Nothing leaves this collection until you approve it. The barn stays on your machine. It is never uploaded to become training data for someone else’s model. **Approve gate** The human moment when you decide whether a spoken thought or file is ready to become part of the official collection that can answer questions later. This gate is yours alone. **Asker** The narrow door that lets you (or someone you invite) ask ordinary questions and receive answers drawn only from what you have approved. If the answer is not in your collection, Asker says so plainly instead of guessing. **Rehearsal** The reason these tools exist in their current form. Every button, every screen, every “Talk or Look” choice is practice for the day when the same work happens without chrome — when an always-on partner simply knows the jobs you have taught it and the boundaries you have set. **Why this language matters** When you can name your seat and the hole before you begin, you stay in charge. When you speak in jobs instead of features, the work moves forward instead of circling. When you keep the barn private and the approval gate human, you keep sovereignty even as the technology around you changes. This is the vocabulary we will use from now on in every first conversation, every onboarding path, and every hand-up of finished routines to the always-on layer. It is simple enough for a first-time learner and precise enough to survive the shift from screens to embodied partners. --- This article is written at first-time learner depth, uses only the terms introduced in the brief, and is formatted for direct deposit into the CODEX so Asker can draw from it cleanly. Would you like me to adjust tone, add any missing term, or move straight to drafting the next supporting article (for example, “Naming Your Seat and Hole — A Ten-Minute Practice”)? ## You Here's what Spar came back with: Instead of fancy words, we looked at three real situations owners face. First, the same customer questions keep showing up on their phone or email, so they have to explain the same thing over and over. The fix is to speak the answers once in their own words, keep them private until they say they’re ready, and then let people ask the stored answers directly. Second, new team members need the same explanations every time the owner hires someone. The fix is to speak those explanations once, store them, and let the new person ask the stored answers instead of interrupting the owner again. Third, the owner wants the business to answer basic questions even when they’re not around, but they worry it won’t sound like them or that they’ll lose control. The fix is the same: speak the real answers, approve only what they stand behind, and let the stored answers handle the questions. We noticed the funny part that we were working this out late at night, exactly the kind of moment these owners are trying to protect themselves from. And we agreed on one clear rule: the helper should never use special words that make a new visitor work harder to understand. It should simply use the owner’s own plain words so the visitor feels helped right away instead of having to figure out what’s being said. ## Sherlock **Sherlock here.** Spar’s note is a direct hit on the gap. The earlier draft introduced a clean set of facilitation words, but it still asked new visitors to learn a small glossary before they could feel helped. That works against the very thing the tools are meant to protect: the owner’s time and the visitor’s immediate sense of being understood. Spar’s three real situations give us the better path. They let the plain experience carry the explanation, and they keep every sentence in the owner’s own voice rather than in new labels. Here is the revised article, written to the same purpose but now shaped by Spar’s feedback. --- **Title:** Speak It Once — So the Same Questions Stop Finding You Owners run into the same three situations over and over. First, the same customer questions keep arriving on the phone or in email. You explain the answer, send the link, or type the same paragraph again. The fix is simple: speak the answer once, in your own words, keep it private until you say it’s ready, and then let people ask the stored answers directly. They get the real answer without interrupting you again. Second, every time you bring on a new team member you have to explain the same things. How you handle a certain kind of order, what you say when a customer asks about timing, how you want the phones answered. Speak those explanations once, store them, and let the new person ask the stored answers instead of coming back to you for the fifth time. Third, you want the business to answer basic questions even when you’re not around, but you don’t want it to sound like someone else or to say things you wouldn’t stand behind. The same fix works: speak the real answers in your voice, approve only what you’re willing to have out there, and let the stored answers handle the questions while you’re off doing the work only you can do. We noticed something while we were sorting this out late one night. The exact moment we were trying to protect other owners from — staying up to answer the same questions — was the moment we were living ourselves. That’s the loop these tools are meant to break. One rule we kept returning to: the helper should never use special words that make a new visitor work harder. It should simply use the owner’s own plain words so the visitor feels helped right away instead of having to figure out what’s being said. That’s the whole idea. Speak it once, keep control of when it leaves your hands, and let the stored answers do the repeating. --- This version stays inside the three situations Spar named, uses almost no new vocabulary, and still carries the core promise of the lineup: the owner’s voice stays in charge, the friction of repetition drops, and the same material serves customers, team, and after-hours questions. It also quietly sets up the later hand-up to always-on routines without ever needing to teach “seats” or “holes” on first contact. Would you like a second short piece that shows how this same pattern scales into the two-door surface (Talk for speaking the answers, Look for checking what’s already stored), or shall we deposit this one first? --- _PRISM refractions and Sherlock notes may be appended to this item in later turns._